The Polish Academy of Sciences in Warsaw. After Poland joined the European Union, its scientists published fewer papers with co-authors in other countries.


Europe’s paradox: Why increased scientific mobility has not led to more international collaborations

A new analysis suggests that lowering barriers to scientific migration can, paradoxically, decrease international collaborations. When top researchers in Eastern Europe started joining high-power institutions in the West, the research suggests, their colleagues and students back home ended up with fewer cross-border connections. "I'm quite surprised how big the effect is," says Paul Nightingale, a science policy expert at the University of Sussex in the United Kingdom who was not involved in the study. "These are worrying findings."

The analysis is based on a natural experiment. In 2004, the European Union expanded by adding 10 member states, including Poland, Hungary, and the Czech Republic. This made it much easier for people in these countries to find work and live in Western Europe. By tapping a government database on intra-European job moves, Alexander Petersen, a computational social scientist at the University of California, Merced, and his colleagues were able to infer the impact of the union's expansion on scientific mobility. The database records the international transfers of highly skilled workers, such as doctors and civil engineers, who need a professional certification in a new host country.

It was easy to see the brain drain. Although most professionals hopped between neighboring countries, the percentage that left Eastern for Western Europe rose from 5% to 29% after 2005. Scientists aren't in the database, because they don't need government accreditation, but Petersen thinks the trend is likely the same. Other experts agree: "There is a lot of migration because of the high-quality infrastructure in universities in the West," says Peter Tindemans, secretary-general of EuroScience, a membership association of European researchers, based in Strasbourg, France. "Scientists gravitate to the places where they can do the best science." He says there is a similar pattern with European Research Council grants, which researchers can take along when they head to a new institution.

To find out what happens when a researcher packs up for a new country, Petersen and his colleagues dug into a database of millions of scientific papers published between 1996 and 2012, looking at affiliations of the authors as a proxy for international research collaborations. After 2004, the new member nations in the east experienced a drop in the proportion of papers published with international co-authors (see chart), the team reports in the current issue of Science Advances. In contrast, this ratio increased in the older member nations. (Here's how the ratio was calculated: Papers were counted as an international publication in each country where a co-author was located. A paper with authors only from a single country was counted as a noninternational publication in that country.) In Poland, for example, the ratio of internationally co-authored papers fell from 33% in 2004 to 28% in 2011, whereas it rose from 28% to 43% in the United Kingdom.

Countries within Western Europe are publishing a higher proportion of papers with foreign co-authors, partly because new talent immigrated from Eastern Europe and brought existing networks of international collaborators.

What likely happened is "rather intuitive," the researchers write in their paper. When top scientists in Eastern Europe moved to institutions in Western Europe, they brought along their connections with scientists in other countries—say, a Hungarian molecular biologist with ongoing collaborations in France and Germany, who moves to the United Kingdom. "Good scientists tend to collaborate with other scientists abroad," says co-author Fabio Pammolli, an economist at the Polytechnic University of Milan in Italy. "This is part of the increasing integration of science." In addition, he notes, talented graduate students came to study in Western universities and built new contacts while working in internationally staffed labs. Meanwhile, their home countries had lost international connections.

This migration and concentration of talent in research hubs is a natural part of scientific growth, the authors say. At the same time, they acknowledge that losing talent is bad for the home country. The tricky balance for policymakers is that the European Union wants its most gifted scientists to be able to chase opportunities and work in the best places, but it also wants equitable growth and development among its member nations—and having a strong science base is a proven part of economic development. "It's one of these traditional European problems," Tindemans says. "We will not find a clear answer."

The European Union has tried to level the economic playing field by providing billions of euros as so-called Structural Funds to its less developed member nations. In the beginning, many funds were not spent well, Tindemans says, and the European Commission has since pushed for better accounting and for more of the funds to be invested in science and innovation. "If you spend it wisely, you could do great things with that money," Tindemans says, pointing to new high-power laser facilities in Eastern Europe that are being built largely with Structural Funds. But he says it will continue to be difficult for Eastern nations to compete with the advanced centers of research elsewhere. "It's impossible that you will have the same high-quality research universities all over Europe. There are too few people and not enough money." 

What is needed in the short term, the authors say, is more incentives for scientific talent to return to Eastern Europe after training and experience in the West. Petersen says the commission already has valuable programs that match scientists in Eastern Europe with Western institutions to build expertise, and fellowship programs for scientific mobility, such as the Marie Skłodowska-Curie actions, include a  mandatory 1-year return period.